R: [PATCH] finish iso support retry

[Giampaolo Tomassoni]

The kernel patch supplying the ATMSAR module does not depend on the evaluation driver.

It is exactly the contrary...

The chain is:

[ATM] <- [ATMSAR] <- [pulsaradsl-1.0.1] <- [libpulsar_fw.a] <- [hardware]

SAR is a thing that has to do with cell encoding and decoding in an ATM vcc context. It is not tied to hardware.

So, ATMSAR actually is pure GPL code. The pulsaradsl-1.0.1 EVALUATION driver is not (see the MODULE_LICENSE() at the end of pulsar.c in the driver). By the way, the purpose of the ATMSAR driver is exactly to be hardware-independant so that you can use it with a PCI as well as with a USB device.

Do I miss something? You're not the first saying this.

Regards,

	giampaolo

> -----Messaggio originale-----
> Da: matthieu castet [mailto:castet.matthieu at free.fr]
> Inviato: martedi 6 settembre 2005 11.21
> A: Giampaolo Tomassoni
> Cc: usbatm at lists.infradead.org
> Oggetto: Re: [ATMSAR] Request for review - update #1
> 
> 
> Hi,
> 
> Giampaolo Tomassoni wrote:
> > Dears,
> > 
> 
> > Also, in order to allow you to evaluate it, I prepared a driver 
> adopting the ATMSAR API for the Globespan Pulsar ADSL PCI card. 
> >This driver is made after the one from Guy Ellis (see 
> http://adsl4linux.no-ip.org/pulsar.html for further reference). 
> >My version is a two-part driver: a GPL one ( 
> http://www.tomassoni.biz/download/pulsar/pulsaradsl-1.0.1-source.t
> ar.bz2 ) 
> >and the same proprietary library you may find in the Ellis' 
> driver from the above-mentioned URL.
> > 
> How could you mix GPL and proprietary library ?
> 
> In your ATM SAR lib you use EXPORT_SYMBOL_GPL. So how this proprietary 
> module could use it ?
> 
> Otherwise I agree it could be a good idea to split atm and usb code from 
> usb atm.
> 
> 
> Matthieu